|Interview by David Bradley
Interview with Peter Loew
Peter Loew was born and grew up in Munich, Germany. He gained a degree in chemistry in 1979 and received his PhD in Organic Chemistry in 1983 from the Ludwig-Maximilian University in Munich. From 1983?1986 he took a Post Doc position at the Technical University in Munich in Computer-Chemistry, which was sponsored by ICI Organics Division, UK.
In 1986, together with Heinz Saller and Johnny Gasteiger, he founded ChemoData GmbH, and in 1989, he and Heinz Saller formed InfoChem GmbH. In 1991, Springer-Verlag (now Springer Science and Business Media) acquired a majority share in both companies and merged them to form InfoChem GmbH. Peter is managing director of InfoChem and still owns shares in the company.
InfoChem focuses on the production and marketing of new chemical information products, and, in particular, on the production of chemical structure and reaction databases and the development of all kinds of software tools required for these applications. The main customers of InfoChem are the big chemical and pharmaceutical companies worldwide, as well as high potential start-ups, and also major publishers such as CAS and Thomson Dialog. InfoChem is also a collaborator of Reactive Reports' publisher ACD/Labs.
There have been a lot of changes in the IT world since InfoChem was formed in 1989. What do you consider the highlights?
From InfoChem's point of view, the most important changes have been the enormous increases in computer speed and memory size combined with a tremendous fall in prices. We can now employ affordable computer hardware where we, or our customers, once had to spend hundreds of thousands of dollars. These developments have enabled InfoChem to process large numbers of chemical structures and reactions in a few hours instead of several months, and to search millions of structures and reactions in a matter of seconds. We could not have developed so many new, innovative chemical algorithms without these changes.
Another significant change has been the evolution of the Internet, which now gives our users easy access to our database products. The organizational effort to cope with this kind of business is considerably less than with other distribution channels. This allows us to offer products we couldn't have dreamt of when we started the company. Last but not least, consider the progress in programming languages and software design techniques. Without these developments it would have been impossible to build and maintain the huge, complex software systems which are part of our products today.
How did the Springer deal come about in 1991?
This is a long story, I'll try to shorten it. In 1988, InfoChem acquired the rights to the electronic SPRESI structural and reaction database. Springer-Verlag was the publisher of the printed Beilstein handbook. When, in 1984, the German government started to fund a huge chemical information project to create the electronic form of Beilstein, Springer-Verlag found itself contributing a large annual payment. As the data structures of SPRESI and Beilstein are quite similar, Springer acquired a majority share in InfoChem with the idea of contributing electronic data instead of money to the Beilstein project. In the end, that idea never came to fruition, but InfoChem has continued as a successful Springer subsidiary.
I see from the InfoChem site that you offer information on 3.6 million reactions and 4.5 million compounds. How does this compare with other proprietary offerings?
The SPRESI data collection is the third largest structure and reaction database worldwide, not far behind CAS and Beilstein. Starting in the 1970s, this database was jointly built by the All-Union Institute of Scientific and Technical Information of the Academy of Sciences of the USSR (VINITI) and the German Zentrale Informationsverarbeitung Chemie, Berlin (ZIC). SPRESI was developed to counteract the Western activities at CAS and Beilstein and, therefore, the data were abstracted from the same sources and with the same quality and comprehensiveness as for CAS and Beilstein. We're pleased that CAS has included the SPRESI reactions in CASREACT since 2001 and, as a result, CASREACT dates back to 1974. We may not have quite as many reactions, but our own Internet offering, SPRESIweb offers search features not available in CAS or Beilstein. For example, SPRESIweb has Name Reactions and Reaction Type searching.
What makes your products unique?
All our products are unique! They all benefit from our enormous know-how in the area of chemical structure and reaction handling that we have accumulated over twenty years in this business. Thanks to the SPRESI data, we were forced, from the very beginning, to work with millions of chemical structures and reactions and all our technology and software developments were always geared to accomplishing exceptional registration and retrieval performance on very large data sets. Our chemistry cartridge for Oracle (ICCARTRIDGE) is the product that combines most of our concepts and developments. The combination of performance and functionality of InfoChem's search engine (ICFSE) is second to none on the market and is one of the key features that makes this Oracle cartridge unique for the users (even if they often don't know that they are using it!). Another extremely important tool is our reaction classification module CLASSIFY. This allows our customers to structure different reaction databases hierarchically, and to link them in products such as our own Global Major Reference Works.
How important is it that resources such as SPRESI are available via the web rather than via solid-state storage or on local intranets?
SPRESI is available both on the web and in-house. The web application is mainly used by small companies and universities while the in-house installations are preferred by larger companies. In fact, InfoChem is the only source that provides such a large database of structures and reactions in computer readable form in SDfile/RDfile format for in-house use.
What is the purpose of your "chemical network"?http://www.infochem.de/eng/texte/references_cooperations.htm?
We implemented this functionality so our users could have easy access to additional information related to the structures and reactions, or references, they are viewing in SPRESIweb. From a technical point of view, we intend to interlink our online databases with other providers of chemical content using the actual structure/reaction or the citation as a crossover key. At the moment we are enriching SPRESIweb with the catalogue data of the big providers of chemicals. Other vendors do similar things, but InfoChem provides a direct correlation between the substances in reactions and their commercial availability.
What is the aim of Global Major Reference Works?
The design and development of electronic Major Reference Works is one of InfoChem's most sophisticated projects. We're not just producing another electronic version of print products; in our approach, these publications become searchable by structure and substructure. This provides tremendous possibilities: it puts an extensive amount of validated information at the user's fingertips, in a highly condensed form. Moreover, in InfoChem's global MRW product we combine several MRWs in one application and provide global search options. In the long term, we hope to forge links between these comprehensive, detailed, and validated secondary sources and the big structure and reaction databases such as CAS, Beilstein, and SPRESI. This would open up completely new ways of instantly gathering chemical information: one can currently only dream of the possibilities.
Do you work with all the major publishers?
Yes, we do work with the major publishers. I've already mentioned CASREACT, and the InfoChem search technology is implemented in a Dialog online service that gives access to about ten million chemical structures. When we started our electronic Major Reference Works projects jointly with several big commercial publishers years ago we also had intensive discussions with both ACS and RSC on the production of similar MRW products and we even did small prototypes for ACS and RSC. Unfortunately, these society publishers did not immediately adopt our ideas, but we still hope they will come round one day.
How did the partnership with Reactive Reports' publisher ACD/Labs come about?
This arose from a customer request that fits very well with our philosophy. We were asked to provide the possibility of calculating predictions of chemical and physical constants and so we decided to partner with ACD/Labs who have a lot more expertise in that area. In return, the user is offered a hyperlink to SPRESIweb when working in ACD/Labs Online (I-Lab).
How do you cope with your diversity of customers from large pharma right down to small start-ups?
Most of our contacts to large pharma are based on personal contacts and relations established over the last twenty years. Large pharma customers usually order licenses and application development in the range of 25,000 euros, or more, a year. Obviously, we try to treat these, our biggest customers, individually. At the other extreme are our many individual or small company customers who cannot afford to pay for personal service or large in-house systems. For them, we distribute SPRESIweb on the Internet. This reduces our overheads and we are able to give them low-price access to a quality product.
In what ways will your and other commercial systems fit with the likes of public access systems such as PubChem?
I see a fundamental difference between public access systems and commercial data sources such as CAS, Beilstein, or SPRESI. The latter offer validated and manually abstracted information that cannot be provided free of charge, especially if quality matters. Public sources, on the other hand, provide an overview and the access to freely available data. In our view it's not an "either...or" but a "both...and". We should push for intensive interlinking and integration of all these different sources, proprietary or public, to provide the users with the optimum system.
Finally, how can traditional commercial organizations survive in a changing market driven by the open-access movement?
This is certainly one of the most challenging developments in the future. From our perspective, OA will play an increasing role in the area of primary publications and thus will create a growing pressure on the well-known publishing houses. It's their turn to react constructively to this movement, as our majority owner Springer Science and Business Media did. Springer is already offering an alternative OA publishing model for authors. However, database building based on an abstracting process and including the necessary quality assurance to maintain a reliable quality standard?be it large databases or selective databases?requires a certain financial effort that has to be recovered by any commercial organization. Therefore, I doubt that high-quality databases will be made available in the OA environment. I believe there will be a co-existence between ever-increasing OA activities and commercial database vendors providing the kind of information that's also indispensable for chemical research.